Considerations for Deployments Using Code Page 932 (Shift-JIS)

The following are special considerations for Japanese-language deployments using Shift-JIS, which is also known as Code Page 932 (or 936 on IBM DB2, or JA16SJIS on Oracle Database). It is assumed that the client computer uses the same code page as the database.

• Some Japanese characters require two bytes per character. However, note that user input can include both single-byte and double-byte chars, in combination. Because there is no direct correlation between length limit and byte limit in the case of double-byte languages, it is impossible to provide direct validation of byte limits. For validation purposes, the assumption should be that all characters typed are double-byte. This may leave unused fields when only single-byte characters are used. Custom validation based on analysis of the input string might be able to solve this problem, if a customer chooses to implement this.

• If your deployment is migrating from Code Page 932 (Shift-JIS) to Unicode, then you must be careful with characters entered on a system where the data is stored in Code Page 932.

  Over 300 characters present in Code Page 932 have only one representation in Unicode, so when these characters are moved to a Unicode system, they are converted permanently to the new value used in Unicode. Because of this conversion, users will see that the character they originally entered has been slightly changed, but the meaning should be the same as before.

  An example is the replacement of the WAVE DASH character by the FULL WIDTH TILDE character, often used in expressing appointment times. There is no correction for this situation, because it is a result of the design of both Code Page 932 and Unicode character sets, which are industry standards.